Published: Nov 19, 2026

Work Analysis

How to Reduce Cycle Time Variation in Manufacturing

    How to Reduce Cycle Time Variation in Manufacturing

    When the same task takes longer than usual at times, the cause may be not only the worker’s proficiency but also component search, re-setting of fixtures, waiting for equipment, etc. Average time alone may miss these delays.

    This article explains the measurement conditions, how to read average, median, and range, and the steps for cause analysis and improvement using the example of automobile part unit assembly. The goal is not speed alone, but to establish work methods where delays and rework are less likely.

    8 Steps to Reduce Work Time Variation

    1. Establish comparison conditions before measuring variation

    Conditions to establish

    Handling when conditions differ

    Part number and component specification

    Separate by type, record condition differences

    Fixture, tool, and setting

    Distinguish as a comparison for different equipment conditions

    Part supply and placement

    Leave differences in shortage, replenishment, and movement as potential causes

    Start and end boundaries

    Re-measure the same event

    Scope of inspection and confirmation

    Check for omitted work

    Time zone and accompanying work

    Record whether replenishment, exchange, and stop are included

    When comparing seasoned workers and newcomers, if conditions differ, not all time difference can be called skill difference. First confirm equipment, materials, and methods, then examine differences in judgment and action.

    2. Break a cycle into work elements and measure each one

    2. Divide a cycle into elements for measurement

    For example, divide into component search, fixture set, fastening, inspection confirmation, and retrieval. Align the boundaries of each section and the cycle boundary. Link video timestamps and measurement records to confirm long sections later.

    3. Use average, median, and range together

    Target

    Calculation and values

    What can be read

    Average

    Total time averaged: 166 ÷ 5 = 33.2 seconds

    Overall time averaged value

    Median

    Smallest order median = 31 seconds

    Approximate median for cycles near center

    Range

    43 − 30 = 13 seconds

    Difference between observed maximum and minimum

    Re-watch the video of the 43-second instance and confirm work not found in normal instances. This can’t prove overall stability with just these 5 instances.

    4. Keep long cycles and retain the cause

    Assume that the 43-second cycle includes 12 seconds for component search. Subtract the search time from this cycle to get 31 seconds, close to the average of the remaining 4 cycles at 30.75 seconds. A hypothesis that search time is related to time delay can be made, but confirm other condition differences before determining the cause.

    Suppose a 43-second cycle included 12 seconds of searching for a part. Subtracting the search leaves 31 seconds, close to the other four cycles’ mean of 30.75 seconds. This suggests that searching contributed to the delay. Check for other differences in the operating conditions before assigning the cause.

    It is possible to split the standard cycle to consider the standard method. However, the average of the overall performance and the occurrence rate of the search should be retained, with the reason for exclusion clearly stated. Long values cannot be discarded as "stable."

    Potential Causes

    Confirm through video and recording.

    Direction for Improvement Proposals

    Search for components and tools.

    Where do you get lost, and what is difficult to distinguish?

    Identification, fixed position, replenishment rules.

    Restart the set.

    Sitting position, component orientation, guide condition.

    Review of guide, guide, and procedure.

    Confusing during display confirmation.

    Viewing position, meaning of display.

    Improve display position and explanation of judgment.

    Equipment and waiting for supply.

    Stop record, replenishment time, upstream process.

    Cause measures for equipment and supply side.

    Judgment varies by person.

    Same scene observed.

    Shared criteria for normal and abnormal judgment.

    5. Distinguish frequency and impact.

    The size of the challenge is seen in both the length of one delay and the frequency of occurrence. For example, if a 12-second search occurs 10 times in 100 cycles, the total is 120 seconds, with an average delay of 1.2 seconds per item. Decide which challenges to address first, considering quality and workload impacts.

    6. Listen to the reason while observing the same section.

    Even if the video appears to be stopped, the individual may be making an important judgment. Confirm "What are you looking at?" and "What conditions change the judgment," linking visible actions to necessary judgments.

    7. Decide which method, layout, or education to change.

    For component identification issues, change layout or display; for sitting issues, change guide and set method; for judgment issues, review standard and materials. Multiple changes should not be combined at once, as this can obscure the reason for effectiveness. Therefore, define the testing scope and confirmation items.

    8. Re-measure under the same conditions.

    Even after improvement, monitor not only the average but also long cycles, number of searches, restarts, confirmation omissions, and workload. Clearly state differences in data variety and observation counts, and do not treat short, accidental values as results. Reflect the validated method as a standard and in education.

    Extract repeating cycles from long-duration video and compare time differences. Ollo Factory’s cycle detection and comparison analysis can be confirmed in product documentation.

    Frequently Asked Questions

    Is it sufficient to educate only individuals with high variability?

    First, confirm differences in supply, equipment, guide, and conditions. If skill or judgment is the cause, avoid simply telling them "work faster." Instead, share the differentiating elements and confirmation methods, then re-evaluate.

    Does a small difference between maximum and minimum values indicate stability?

    The range is affected by observation count and the impact of single-value chance. Use combined data distribution, occurrence order, and cause recording under consistent conditions to determine which conditions produce long values.

    Summary: Confirm reasons for long cycles and link to reproducible improvements.

    • Align part number, equipment, supply method, and measurement range.

    • Use average, median, and range together, recording reasons without eliminating long segments.

    • Test improvement proposals in a single process or limited part number, confirming reduced variability under the same conditions.

    Comparing multiple cycles manually creates a burden, which can be reduced with AI. Automate cycle detection and time comparison of the response area, making it easier to return to situations where differences arise and check conditions.

    Ollo Factory, a manufacturing-specialized work analysis AI that can automate cycle detection and work time comparison.

    ① Arrange the same work area for comparison.

    Compare work videos and identify elements with time differences. Return to the movement of the set, component search, binding, and inspection areas, and confirm what is different in the video.

    Screen in Ollo Factory to compare the operation and time difference of the same work area videos.

    Official demo screen of Ollo Factory. This is a different task from the unit assembly example in the text.

    ② Track cycles with differences.

    Compare work times for each cycle and view videos of the longer ones. Investigate component searches and redo work that may be overlooked with the average value alone, helping to determine which work method, arrangement, or education to review.

    Screen in Ollo Factory to confirm the time difference of each cycle for repetitive work with a graph.

    Official demo screen of Ollo Factory. This is a different task from the unit assembly example in the text.

    Differences alone do not confirm the cause. Do not evaluate only the worker's speed; align conditions such as equipment, variety, and component supply to confirm.

    Watch Ollo Factory’s Work Comparison AI in Action

    About the author

    K.Y. | Business Development Lead & Marketing, Ollo Inc.

    K.Y. leads business development and marketing at Ollo Inc., an AI startup originating from the Matsuo Laboratory at the University of Tokyo and the developer of Ollo Factory, a work analysis AI for manufacturing.
    After studying engineering and design at university, K.Y. joined a major automaker and worked on new electric-vehicle development, from design through factory production, including hands-on assembly-line training. K.Y. later worked in sales and marketing at a manufacturing startup.
    Today, K.Y. helps manufacturing teams use AI to improve the way they work through Ollo Factory.

    Back to articles

    Related articles

      See how Ollo Factory can support your factory

      Book a demo

      • See the product in action
      • Discuss your production needs
      • Explore available plans

      Book a demo

      Request a brochure

      • Customer stories
      • Key features
      • How to get started

      Ollo Factory English product brochure